Available online at www.sciencedirect.

com

CERAMICS
INTERNATIONAL
Ceramics International 41 (2015) 8904–8914
www.elsevier.com/locate/ceramint

Microstructural, mechanical and tribological properties of nanostructured

YSZ coatings produced with different APS process parameters
C. Lamutaa, G. Di Girolamob,n, L. Pagnottaa
a
Department of Mechanical Engineering, University of Calabria, Ponte P. Bucci, Cubo 44C, Rende, Italy
b
ENEA, Materials Technology Unit, Casaccia Research Centre, Rome, Italy
Received 15 January 2015; received in revised form 9 March 2015; accepted 23 March 2015
Available online 30 March 2015

Abstract

Plasma sprayed ceramic coatings can be used in turbine engines as thermal barrier or abradable coatings, in order to improve the durability of
the components as well as the efficiency. The presence of nanostructures, deriving from partial melting of agglomerated nanostructured particles,
represents an interesting technological solution in order to improve their functional characteristics. In this work nanostructured yttria stabilized
zirconia (YSZ) coatings were deposited by air plasma spraying (APS). The influence of the main process parameters on their microstructural,
mechanical and tribological properties was investigated by scanning electron microscopy (SEM), indentation techniques at micro- and nano-scale
and wear tests, respectively. Their porous microstructure was composed of well melted overlapped splats and partially melted nanostructured
areas. This bimodal microstructure led to a bimodal distribution of the mechanical properties. An increase of plasma power and spraying distance
was able to produce denser coatings, with lower content of embedded nanostructures, which exhibited higher elastic modulus and hardness as
well as lower wear rate.
& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: C. Hardness; C. Wear resistance; Plasma spraying; Young's modulus; Zirconia

1. Introduction Ceramic materials with low thermal conductivity and heat

Ceramic coatings are suitable to be employed for thermal
protection of Ni-based superalloy turbine components operat-
ing in power plants and aircraft engines. Their application
allows improving their high-temperature capability and dur-
ability, by reducing the heat flux and the temperature at the
metal surface. Significant improvements in terms of engine
efficiency and lower pollution are then expected [1,2].
A thermal barrier coating (TBC) system is usually composed
of a metal substrate, a metallic bond coat and a ceramic top
coat [3]. The intermediate bond coat (MCrAlY) plays a
meaningful role on the adhesion of the ceramic top coat and
provides better resistance to the attack typically promoted by
oxygen and molten salts in severe working environments [4].
n
Corresponding author.
E-mail address: giovanni.digirolamo@enea.it (G. Di Girolamo).
capacity are good TBC candidates.
Partially-yttria stabilized zirconia (8YSZ) is the most
common used TBC material, owing to its satisfactory thermal
and mechanical properties (low thermal conductivity, relatively
high thermal expansion coefficient, low Young's modulus,
high hardness and toughness) [5,6].
Plasma spraying is suitable for fabrication of thick porous
coatings on complex metal parts. In such process powder
particles are injected in the plasma jet by an inert gas, melted
and accelerated toward the substrate, where they impact at high
speed and quench, thus producing the build-up of a coating
with typical microstructural defects such as splat boundaries,
pores and microcracks [7].
During service at high temperature the infiltration of oxygen
through the microcracks and interconnected pinholes of the
ceramic TBC and the outward diffusion of bond coat con-
stituents, such as Al, Cr and Ni, lead to the oxidation of the
bond coat surface and to the following formation of the TGO

http://dx.doi.org/10.1016/j.ceramint.2015.03.148
0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914
(thermally grown oxide) at the interface. The gradual growth
of the TGO layer assists the evolution of tensile stresses and
the growth of horizontal cracks at the interface, thus leading to
TBC spallation and to the failure of the whole TBC system
[8,9]. To this purpose it has been reported that the use of
nanostructured materials is able to promote the formation of a
thin, uniform and dense TGO, partially suppressing the fast
growth of other mixed oxides, reducing the oxidation rate and
prolonging the lifetime of TBC systems exposed at 1000 1C
and 1150 1C under cycling conditions [10,11]. Indeed, the
mixed oxides are usually characterized by high brittleness and
microcrack growth.
Concerning ceramic TBCs it has been reported that sig-
nificant enhancements can be achieved by using nanostruc-
tured materials in substitution of conventional ones. Indeed,
the reduction of the grain size typically involves better
mechanical strength and toughness [12,13]. The nanostruc-
tured coatings are characterized by higher interlamellar
strength, i.e. the lamellae are well bonded each other in
comparison with the conventional coatings owing to better
filling and compactness between at splat boundary, as well as
by higher crack propagation resistance, because the nanozones
typically embedded within the microstructure act as crack
arresters. The high porosity of nanostructured coatings also
involves better thermal properties and thermal shock charac-
teristics [12].
It is worth noting that single nanoparticles cannot be carried
by plasma jet and deposited on the substrate, so that they are
commonly pre-synthesized in micronsized particle agglomer-
ates. To this purpose, these agglomerates should be only
partially melted to preserve part of their starting nanostructure.
On the contrary, if the agglomerates are totally melted, the
grain growth occurs and the final microstructure resembles that
of a conventional coating [14].
In the present work different process parameters were
employed for coating manufacturing in order to control the
degree of melting of the powder particles and to obtain
coatings with well-tailored characteristics. The morphology
and the microstructure of nanostructured YSZ coatings were
investigated by scanning electron microscopy (SEM). Nano-
and Micro-Indentation (NI, MI) tests were employed to study
the evolution of the main mechanical properties, such as
Young's modulus and hardness, whereas the wear rate was
obtained by means of a ball-on-disk test. A statistical approach
was used to study the mechanical properties of YSZ coatings
and their relationship with the microstructure.
2. Experimental procedure
2.1. Plasma spraying
YSZ ceramic coatings were deposited on Ni superalloy
disks (IN738, ϕ ¼ 25 mm, thickness ¼ 4 mm). The substrates
were sand blasted using an alumina abrasive powder to
increase their surface roughness and to improve the mechanical
interlocking between coating and substrate. The substrate
roughness, measured using a three dimensional optical surface
8905
profilometer, was found to be 6.9 7 1.1 μm. An atmospheric
plasma spraying equipment, with a 4F-MB plasma torch with a
6 mm internal diameter nozzle, was used for coating deposi-
tion. A metallic CoNiCrAlY coating (Amdry 995C, Sulzer
Metco) with thickness of 150 μm was previously applied as
bond coat on the substrate surface.
Nanostructured partially yttria stabilized zirconia TBCs
were then deposited using the nanostructured ZrO2–7 wt%
Y2O3 powder feedstock (Nanox S4007, Inframat, US). The
final thickness of the coatings was of about 300 μm whereas
their roughness was equal to 8.3 7 1.0 μm (no significant
changes were appreciated for the samples produced using
different process parameters). Two sets of six different samples
were produced. The six samples of each set differ in the
process parameters employed: the arc current was varied on
three levels (500 A, 565 A and 630 A) and the substrate–torch
distance on two levels (80 mm and 100 mm). The other
parameters were kept constant and can be summarized as
follows: primary gas flow rate (Ar) 40 slpm, secondary gas
flow rate (H2) 12 slpm, powder flow rate 28.5 g/min, and
substrate tangential speed 2086 mm/s.
2.2. Microstructure
The phase composition of nanostructured zirconia powder
and coatings produced using different processing parameters
was investigated by an x-ray powder diffractometer (XRD PW
1880, Philips, Almelo, Netherlands) operating with CuKα
radiation (λ ¼ 0.154186 nm) produced at 40 kV and 40 mA.
The analyzed range of the diffraction angle 2θ was between
201 and 801, by a step width of 0.021 and a time per step equal
to 5 s.
The morphology and the microstructure of powder feedstock
and as-sprayed YSZ coatings were analyzed by scanning
electron microscopy (SEM-LEO 438 VP, Carl Zeiss AG,
Oberkochen, Germany). The SEM pictures were then pro-
cessed by image analysis software (Image J, U.S. National
Institutes of Health, Bethesda, MD, USA) to measure the
percentage of molten and semi-molten areas embedded in
coating microstructure, and the distribution of the nanostruc-
tured areas. The size of the regions used for porosity
measurements was 350  250 μm2.
2.3. Mechanical properties
The mechanical properties of YSZ coatings were determined
by Micro- and Nano-Indentation tests. A measuring system of
CSM Instruments SA, Peseux, Switzerland, equipped with
three objective lenses (with magnitude of 5  , 20  and
100  ) and a ConScan, was used.
In order to perform indentation tests the cross sections of the
first set of coated samples were prepared using a standard
metallographic procedure for ceramic coatings, including low-
speed sectioning, cold mounting in vacuum in two-part epoxy
resin, grinding, polishing and finishing to 0.25 μm.
The indentations were performed on a portion of the ceramic
top coat containing areas with different melting degrees and
8906 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914
were equally spaced (40 μm for NI and 60 μm for MI) in order
to avoid the mutual influence of consecutive indentations [15].
Due to the presence of microstructural defects such as pores
and poorly compacted areas, an Adjust Depth Offset operation
was set every 10 indentations in order to find the height
position of the sample surface (acronym ADO in Fig. 1).
NI tests were performed according to a 4  10 matrix (with
lines parallel to the substrate) by using a Berkovich tip, a
loading and unloading speed of 3 mN/s, a hold time of 10 s
and two different values for the maximum load: 8 and 100 mN.
MIs, distributed according to a 3  10 matrix, were also
performed by using a Vickers microindenter with maximum
loads of 50 gf and 100 gf and a hold time of 10 s.
The values of reduced Young's modulus Er and hardness H
were obtained by load–depth curves, according to Oliver and
Pharr theory [16,17] and then the experimental data were
analyzed by assuming a two-parameters Weibull statistical
distribution [18].
2.4. Tribological properties
The second set of specimens was addressed to wear tests.
For the tribological characterization ball-on-disk tests were
carried out using a CSM Instrument Tribometer. The tests were
performed in dry mode with a sintered α-Al2O3 ball (6 mm in
diameter), a wear track radius of 5 mm, a normal load of 10 N,
a linear speed of 0.1 m/s and a cycle number of 20,000. The
wear rate is calculated according to the following formula:
v using scanning electron microscopy.
WR ¼ ð1Þ
Fl
where WR is the wear rate [mm3/(N m)], v is the worn volume
[mm3], F is the normal applied load [N] and l the sliding
distance [m]. The track profile was determined by means of a
Taylor-Hobson Surtronic 25 profilometer (Rand Taylor Hob-
son Ltd., Leicester, UK) and as the maximum depth of the
worn area exceeded the full scale of the profilometer (300 μm)
the CSM Instrument ConScan was used. Ten randomly located
profile measurements were performed on each tested sample
Fig. 1. Disposition of NI on the top coat cross section. Starting from the top:
epoxy resin, top coat, bond coat and metallic substrate (magnitude 5  ).
after removing wear scraps by an air jet. The wear tracks were
finally analyzed by scanning electron microscopy.
3. Results and discussion
3.1. Microstructure
X-ray diffraction patterns of YSZ powder and coatings
produced using different processing parameters are shown in
Fig. 2: the high-angle (72–751) region of the diffraction
patterns showing (400) zirconia reflections is drawn. The
powder feedstock is mainly composed of tetragonal t zirconia
phase with smaller amounts of cubic and monoclinic zirconia
phases, according to Joint Committee on Powder Diffraction
Standards (JCPDS) available at International Centre for
Diffraction Data (no. 81-1544 for tetragonal, no. 49-1642 for
cubic and no. 37-1484 for monoclinic zirconia). All the
coatings are composed of a mixture of metastable t0 zirconia
phase and tetragonal t phase. The metastable t0 phase is
associated with the molten part of the powder agglomerated
particles, whereas the tetragonal t phase is associated with their
retained nanostructured unmelted nanostructured part. No
monoclinic phase was detected in the coatings. The calculation
of the percentage of t0 and t zirconia phases cannot be properly
performed by XRD spectra because of the overlapping
between the related peaks. As reported below, more reliable
estimation can be obtained by studying the distribution of
retained nanostructured areas in the coating microstructure
Fig. 3 shows the cross sectional SEM microstructure of
nanostructured YSZ coatings produced at the spraying distance
of 80 mm and using different values of plasma current (500,
565 and 630 A). The plasma current is proportional to the
plasma power and to the temperature of the sprayed particles,
so that it can be used to have a good control on the degree of
melting of the particle agglomerates. All the coatings exhibited
a bimodal microstructure composed of well melted splats (dark
Fig. 2. High-angle region of XRD spectra for YSZ (a) powder and coatings
produced using different process parameters: (b) 500 A and 80 mm, (c) 500 A
and 100 mm, (d) 565 A and 80 mm, (e) 565 A and 100 mm, (f) 630 A and
80 mm, and (g) 630 A and 100 mm.
 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914 8907

gray dense areas in the pictures) and partially melted nanos-
tructured areas (light gray areas) possessing an intrinsic
porosity derived from their pre-agglomeration, as shown in
Fig. 4(a). The nanostructured areas are composed of loosely
bound nanostructured particles.
Fig. 4(b) shows the morphology of the related powder particles,
so that it can be deduced that the nanostructured zones retained in
the final coatings were not affected by complete melting during
processing. The inset shows some clusters of nanograined particles
with size close to 100 nm. During processing the hot plasma gas
penetrated inside the agglomerated particles of the powder feed-
stock, melting their surface, while their core remained unmelted.
The low heat transfer associated to their intrinsic porosity, the short
residence time of the same sprayed particles in the plasma jet and
the high quenching rate of the solidified splats at the substrate
surface were able to reduce the mechanism of grain growth and
nucleation, thus preserving great part of the starting nanostructure
in the final coating.
As well displayed in Fig. 3, the well melted areas cement the
loose microstructure, providing good mechanical integrity. The
distribution of the nanostructured areas is more uniform for
lower values of plasma current. For increasing values of this
parameter the melting degree of the powder particles enhances,
thus leading to denser coating with lower content of nanos-
tructured areas. The flight path and the temperature history of
the agglomerated particles in the plasma jet affect their
distribution in the final coatings. The melting process is
strongly related to the temperature distribution in the plasma
jet and to the heat transfer to the porous agglomerates.
Table 1 reports the amount of retained nanostructured areas
measured by image analysis along the cross section of
nanostructured YSZ coatings produced using different process
parameters. At constant spraying distance the percentage
of nanozones decreases with increasing the plasma current,
due to the better melting of the particle agglomerates, so that
denser coatings are produced. This effect is more pronounced

Fig. 3. Cross sectional SEM microstructure of nanostructured YSZ coatings deposited at spraying distance of 80 mm and using various plasma current levels: (a)
500 A, (b) 565 A and (c) 630 A.

Fig. 4. (a) Cross sectional SEM microstructure showing the morphology of the retained nanostructured areas; (b) a view of agglomerated nanoparticles in the
powder feedstock with detail of zirconia clusters.
8908 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914

Table 1 blade tips and the walls of the shroud, in order to provide seal
Fractions (%) of retained nanostructured areas measured in YSZ coatings
produced using different process parameters (plasma current and spraying
and improve the efficiency.
distance). On the contrary, lower retention of nanostructures (20%) is
preferable for development of thermal barrier coatings with
Distance Current low thermal conductivity, high structural integrity and resis-
500 A 565 A 630 A tance to the infiltration of oxygen and molten salts in severe
working environments (stator turbine blades and vanes). The
80 mm 367 4 2371 2074 presence of nanostructured areas with low sintering rate can
100 mm 317 2 2574 2171 counteract the high-temperature densification of YSZ coatings,
which negatively affects the compliance and the thermal
property, thus retarding extended microcracking and TBC
delamination [12].
Nanozones play a significant role on the thermal shock
resistance and durability of TBCs, because they act as crack
arresters in the case of microcracking promoted by thermal
stresses produced by thermal expansion mismatch between
overlapped layers, thus increasing the fracture toughness
[12,19]. For this reason a Vickers indentation at very high
load (5000 gf) was performed on one of the produced samples
in order to induce cracks (white circles in Fig. 5). In the
detailed micrograph taken at higher magnification it can be
observed as cracks that propagate in the molten phase and are
arrested by nanozones.
Columnar grains with diameter in the range from 50 to
300 nm are detectable within coating microstructure (Fig. 6).
Fig. 5. Crack arrest produced by nanozones in plasma sprayed nanostructured
They are oriented along the direction of the grain growth and
YSZ coating.
derived from heterogeneous nucleation at splat boundary
produced by the heat flow released by the crystallization of
the previous deposited splats. Some equiaxed grains can be
also observed at splat boundary. They are produced by
homogeneous nucleation, occurring when the heat loss pro-
moted by cooling at the substrate is higher than the heat
released by the crystallization [19].

3.2. Mechanical properties

Tables 2 and 3 summarize the values of Young's modulus,

Fig. 6. Fractured coating cross section showing the morphology of the melted
splats, composed of columnar and equiaxed grains.
at 80 mm. The effect of spraying distance is more pronounced
at the lowest value of plasma current. Higher spraying distance
involves higher residence time of the sprayed particles in the
plasma jet, better melting and thus higher deposition efficiency
and higher coating thickness.
It is worth noting that the percentage of nanostructured areas
embedded in coating cross section can be properly optimized
based on the application the coating is addressed to. High
retention of nanozones (30–40%) is particularly desired for
manufacturing of abradable coatings with relatively low
mechanical integrity. These coatings can be used in turb-
ine engines to minimize the bypass flow of hot combustion or
cold compressor gases through the spaces between the rotating
nano- and micro-hardness measured on the cross sections of
nanostructured YSZ coatings, deposited at spraying distances
of 80 and 100 mm, respectively, and using different values of
plasma current parameter. The measured values magnitude is
in good agreement with that of results in literature [20,21].
It should be noted that the mechanical properties tend to
increase with increasing the plasma current and the spraying
distance. Specifically, at constant spraying distance, for
increasing values of current, an average increase of 9% and
30% for Er and 13% and 60% for H, was observed, from
nanoscale to microscale, respectively. Instead, the influence of
the spraying distance is less prominent. At constant current
value, for increasing values of torch–substrate distance, an
average increase of 4% and 7% for Er and 3% and 34% for H
was recorded, from nanoscale to microscale. It should also be
noted that the values of the mechanical properties exhibited
large scattering.
Fig. 7 shows the load–depth (P–d) curves obtained by
nanoindentations performed in the melted areas (continuous
line) and in partially melted ones (dashed line). The curves
 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914 8909

refer to coatings produced at 500 A and 100 mm, but similar It is possible to observe that in the melted phase lower
behavior was noticed for all the other samples. The parameters maximum penetration was noticed (709 nm versus 2056 nm in
used for NI were: maximum load 100 mN, loading and the partially melted phase) as well as higher unloading curve
unloading speed 3 mN/s, hold time 10 s. slope. This suggests that the melted phase is characterized by
higher stiffness and hardness in comparison with the partially
Table 2 melted area [16,17].
Mean values and standard deviations of the mechanical properties for YSZ Indeed, as shown in Fig. 8, the size of the imprint is bigger
coatings sprayed at 80 mm and using different values of plasma current,
in partially melted area. The nanostructured areas are char-
measured by Micro- and Nano-Indentation tests using various loads.
acterized by porosity at nanoscale which reduces the cohesion
Load Current (A) Er (GPa) H (GPa) between the agglomerated nanostructured particles. In addi-
tion, the curves in Fig. 9 show that the melted phase reveals a
8 mN (NI) 500 151747 11.175.4
565 170733 11.874.9
weak creep phenomenon, whereas the unmelted one is quite
630 174742 13.774.8 affected by it: in fact, it presents a great depth increase during
100 mN (NI) 500 100737 6.373.9 the hold time at the maximum load. Based on the observations
565 121737 8.374.7 herein reported, one can conclude that the increase of the
630 125741 7.974.3 mechanical properties, observed in Tables 2 and 3, is related to
50 gf (MI) 500 76720 4.272.4
565 114728 6.173.2
the increase of the melted fraction in YSZ coatings when
630 123731 8.775.0 higher values of current and distance are employed (Table 1).
100 gf (MI) 500 80719 3.471.4 The usage of increasing current values involves higher
565 100726 4.472.5 plasma power, higher temperature of the sprayed particles
630 125721 7.272.7 and therefore lowers retention of nanostructured areas and
higher coating density. As previously reported, higher distance
involves higher residence time of the sprayed particles in the
Table 3 plasma jet and better degree of melting, even if the effect of
Mean values and standard deviations of the mechanical properties for YSZ distance on the experimental trend of the mechanical properties
coatings sprayed at 100 mm and using different values of plasma current,
was not always consistent, probably because of a kind of
measured by Micro- and Nano-Indentation tests using various loads.
mutual influence with current. This aspect could be analyzed in
Load Current (A) Er (GPa) H (GPa) future works by performing an experimental plan that provides
a third level of torch–substrate distance and more in-depth
8 mN (NI) 500 1597 68 9.9 75.8
565 1777 54 11.7 75.1
analysis. As clearly observed in Tables 2 and 3, the values of
630 170 7 60 12.176.2 the mechanical properties measured at microscale are lower
100 mN (NI) 500 1087 37 7.1 74.3 than those obtained at nanoscale, because of the larger
565 1247 53 9.1 76.2 analyzed volume and the higher influence of typical defects
630 1357 36 9.5 74.0 embedded in coating microstructure, such as coarse pores,
50 gf (MI) 500 787 40 4.4 73.3
565 1257 40 10.376.3
splat boundaries and microcracks.
630 1587 73 19.6712.9 Moreover, for each indentation technique employed, as the
100 gf (MI) 500 767 20 3.3 71.9 indentation load rises the analyzed volume grows and the same
565 1137 29 5.9 72.9 microstructural defects become more significant, enough to
630 1097 22 5.5 72.6 bring down the mechanical properties, that result always more
averaged and far from those of bulk stabilized zirconia [14,15].
Further analysis showed that the mechanical properties do
not change across coating thickness and their variation is so
small to be covered by the high scattering of the results (related
to the different characteristics between melted and partially
melted phases). In order to understand the reasons of this high
scattering, two series of NIs were performed on the areas with
different morphologies, by using a maximum load of 8 mN
and a time for loading and unloading of 10 s, without any hold
time. Fig. 9 shows some of the load–depth curves obtained.
It is interesting to notice that the curves related to the melted
area (continuous lines) are very close and repeatable, whereas
the curves acquired during indentation in the unmelted area
(dashed lines) are somewhat different and scattered. The well
melted area is, in fact, characterized by lower scattering in the
mechanical properties, whereas the second one exhibits widely
Fig. 7. Load–depth (P–d) curves corresponding to melted and partially melted scattered data (Table 4), typical of structures produced by
phases. nanostructured agglomerates assembly and characterized by
8910 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914

Fig. 8. Indent of NI (Berkovich) in (a) partially melted and (b) melted phase (magnitude 100  ).

assuming a two parameters Weibull distribution for the mech-

anical properties. The data can be approximated with a linear
regression curve. On the same graph, the values of shape and
scale parameters are reported. According to the previous
assertions, the melted phase, characterized by less scattered
data, presents greater values of m. The probability density
functions of Er and H are plotted, respectively, in Fig. 11(a)
and (b), in gray for the melted area and in black for the
partially melted one.
The sample obtained by merging the experimental data related
to melted and unmelted areas can be considered like a coating with
50% of the two phases. The analysis of the distributions concerning
this kind of sample can be very useful for the prediction of the
mechanical properties of any other nanostructured coating. For this
Fig. 9. Load–depth (P–d) curves corresponding to melted and unmelted purpose, in Fig. 12, Er and H Weibull plot of this simulated coating
fractions and measured by NI at 8 mN.
is shown, while Fig. 11(a) and (b) shows (dashed lines) the related
probability density function (obtained as a weighted average of the
Table 4
Elastic modulus and hardness of melted and unmelted phases.
probability density of the data referred to single zones). It can be
noted that the points related to partially melted phase are so
Mechanical properties Melted area Unmelted area scattered that they contaminate the data of the second section,
concerning the melted area, and go also beyond this region (in this
Er (GPa) 196729 131 761
H (GPa) 15.872.3 9.077.8 example it is just a point). The slope of the mixed section (the
second one) increases by decreasing the level of contamination.
Despite the data merging, the bimodal behavior of the coating can
be clearly observed, due to the change in the slope of the Weibull
plot. In order to find the transition point between the two regions
the R2 statistics analysis can be used [14]. It should be noted that
the presence of points above the mixed zone may generate
incorrect interpretations of the data distribution, by suggesting,
erroneously, the existence of a third phase, and then a trimodal
distribution. Fig. 13 shows the trends of the mechanical properties
for the coatings produced in this work and tested by Nano-
Indentation at 8 mN.
In all the cases a bimodal Weibull distribution can be
noticed, as also observed in any previous works [14,20]. The
zone in the graph characterized by lower values of H or Er
Fig. 10. Weibull plot of Er and H for melted and unmelted phases; the shape reflects the mechanical behavior of the partially molten phase
(m) and the scale (ϑ) parameters are also reported.
and it is characterized by lower slope, owing to greater
scattering, whereas the second one is characteristic of the
intrinsic porosity. Table 4 summarizes the mean values and the mixed area. It is interesting to notice that, at constant torch–
standard deviations of the elastic modulus and hardness for substrate distance, as current value rises the slope of the
both these areas. second regression line increases: this is caused by the increase
Fig. 10 shows the distribution of Er and H, in a biloga- of the degree of melting of the sprayed particles which results
rithmic scale [18], for the well melted and unmelted areas, by in lower probability of contamination of the data pertaining to
 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914
Fig. 11. Probability density function of (a) Er and (b) H for melted (gray) and unmelted (black) phases and of the resultant bimodal distribution (dashed).
Fig. 12. Weibull plot of H and Er data for the sample obtained by merging the
data arising from melted and unmelted areas.
the molten zone. Note that, at constant current level, the
increase of the torch–substrate distance produces different
trend, but the results are in good agreement with the volume
fractions of nanostructured areas listed in Table 1.
The results obtained by NI tests performed at maximum load
of 100 mN showed similar behavior, but less noticeable
changes in slope between the interpolating sections were
detected. This effect is the consequence of the greater size of
the volume involved during NI at higher load, which mediates
the local properties and makes the bimodality of the distribu-
tions more difficult to be observed.
In turn the results obtained by Micro-Indentation for the
coatings produced at 500 A and 100 mm are shown in Fig. 14.
The second region of the distributions of MI data exhibits
lower slope with respect to the first one, unlike NI tests, and in
agreement with the data reported in literature [14,20].
This behavior can be addressed to the scale effect for which
microcracks and globular pores embedded in the molten areas
are responsible of data scattering, since higher volume of
material was under analysis [14].
Therefore at higher indentation loads the data distribu-
tion is almost single-mode, as shown in Fig. 14(b). Similar
behaviors were observed for the other five specimens
produced with different values of current and torch–
substrate distance.
8911
3.3. Tribological properties
Fig. 15 shows result about wear rate of the second set of
samples tested by the ball-on-disk test. By comparing the wear
behavior of nanostructured coatings to that of a conventional
one (characterized by the same initial roughness of about 8 μm
and tested with the same parameters listed in Section 2.4) it
can be noted that just the sample with the lowest nanozones
percentage (20%) has a higher wear resistance and that the
wear rate increases with increasing the values of nanozones
percentage (values in Table 1). For specimens with 31% and
36% of nanozones the coating was completely worn until the
appearance of the superalloy substrate during the test.
Many researchers have reported superior wear performances
for nanostructured ceramic coatings when compared to those
of conventional coatings [12]. In this kind of coatings in fact
the wear mechanism is a combination of brittle fracture and
abrasive wear (promoted by delamination and detachment of
small fragments or particles from the surface) and the presence
of dense nanozones limits both of these phenomena. The brittle
fracture is confined by the crack arrest effect whereas the
abrasive wear is reduced by the wear scars of the nanostruc-
tured coating that are finer and smoother than those of the
conventional coatings and form an adherent layer which limits
wear [22].
These anti-wear performances refer to coatings with rela-
tively low nanozones percentage (less than about 20%). When
the amount of nanozones is higher (as observed for the
samples analyzed in this work) the molten part of the semi-
molten agglomerates cannot fully infiltrate into the capillaries
of their non-molten core during thermal spraying by causing
the formation of porous and friable nanozones, typical of
abradable coatings [12].
Therefore by varying process parameters the percentage of
retained nanozones can be controlled in order to produce
abradable coatings with different wear degrees or thermal
barrier coatings with higher mechanical integrity.
Fig. 16 shows the wear tracks of the three samples produced
at torch–substrate distance of 100 mm and using different
current values. It can be observed that the track becomes more
pronounced with decreasing the values of the current and
increasing the percentage of retained nanozones. The dark zone
8912 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914

Fig. 13. Weibull plot of H and Er for nanostructured YSZ coatings fabricated using different process parameters (plasma current and spraying distance) and tested
by NI at 8 mN.

Fig. 14. Weibull plot of H and Er for nanostructured YSZ coating produced at 500 A and 100 mm and tested by MI at (a) 50 gf and ( b) 100 gf.
 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914 8913

in the middle of the track in Fig. 16c) represents the metallic

Fig. 15. Wear rates of nanostructured YSZ coatings (produced with different
process parameters and characterized by different percentages of nanozones
(Table 1)) and conventional YSZ coatings.
substrate by confirming the results shown in Fig. 15. The
morphology of the wear area is the same for all the nanos-
tructured coatings produced and it is shown in Fig. 17. In the
magnification of Fig. 17(b) the white holes highlight the cracks
caused in the molten zones during the wear test that indicate the
occurrence of a brittle fracture. However, the observed cracks
are very few and the dominant wear mechanism of the analyzed
nanostructured coatings, characterized by a high percentage of
friable nanozones, comes from abrasive wear. Indeed, the darker
areas in the worn zone (observable in Figs. 16 and 17) may
represent not only the molten zones, but also the compaction of
the fine debris produced during the wear test.

Fig. 16. Wear tracks of samples produced with a distance of 100 mm and different current values: (a) 630 A (21% of nanozones); (b) 565 A (25% of nanozones);
and (c) 500 A (31% of nanozones).

Fig. 17. (a) Wear zone morphology of the nanostructured coatings: the dashed curve represents the inner edge of the wear track; (b) magnification of the area in the
white box of (a).
8914 C. Lamuta et al. / Ceramics International 41 (2015) 8904–8914
4. Conclusion
In this work the influence of some process parameters
(plasma current and stand-off distance) on the microstructural,
mechanical and tribological properties of plasma sprayed
nanostructured YSZ coatings was investigated. The partial
melting of the nanostructured particle agglomerates led to the
formation of a bimodal microstructure composed of well
melted splats and semimolten areas. Columnar and equiaxed
grains were observed in the melted splats, while the partially
melted areas did not suffer grain growth and retained porosity
at nanoscale. An increase of plasma current promoted the
increase of the particle temperature, improving the melting
degree and the density of the coatings.
The reduction of the retained nanostructured areas produced
significant increases of Young's modulus, hardness and wear
resistance. The effect of the torch–substrate distance was more
significant at lower current values.
The Nano- and Micro-Indentation tests, performed at
various loads, in conjunction with Weibull statistical approach,
revealed that the bimodal microstructure involved a bimodal
distribution of the mechanical properties, which tended to
disappear with increasing the indentation load. The elastic
modulus and the hardness decreased with increasing the
indentation load, because of higher analyzed volume and
stronger influence of the microstructural defects.
The ball-on-disk tests performed revealed that the nanostruc-
tured coatings produced with the APS process parameters set in
this work were mostly abradable coatings characterized by wear
rate higher than that of the conventional coatings, due to high
presence of nanozones. The SEM images of the wear area
showed that the morphology is the same for all the produced
nanostructured coatings and that the dominant wear mechanism
comes from abrasive wear rather than a brittle fracture.
The analysis of the influence of APS process parameters on
coatings properties is really useful in order to engineer
nanostructured YSZ coatings for different applications.
The development of a reliable procedure allowing a proper
control of the percentage of the nanostructured areas embedded
in coating microstructure is not easy to be achieved and will
require a more detailed plan of experiments with the introduc-
tion of other process parameters and further investigations.
Acknowledgments
The nano- and micro-hardness measurements were per-
formed with equipment kindly provided by “Centro Materiali,
Tecnologie e Ricerca Avanzata (MaTeRiA)” from University
of Calabria, funded with “Pon Ricerca e Competitività 2007/
2013, Asse I Sostegno ai Mutamenti Strutturali Obiettivo
Operativo 4.1.1.4, Potenziamento delle Strutture e delle
Dotazioni Scientifiche e Tecnologiche, Azione I - Rafforza-
mento Strutturale”.
The authors wish to thank C. Blasi for her valuable con-
tribution coating manufacturing and P. Caliandro for prelimin-
ary SEM observations.
References
[1] J.A. Nesbitt, Thermal response of various thermal barrier coatings in a
high heat flux rocket engine, Surf. Coat. Technol. 43/44 (1990) 458–469.
[2] N. Curry, N. Markocsan, X.A. Li, A. Tricoire, M. Dorfman, Next
generation thermal barrier coatings for the gas turbine industry, J. Therm.
Spray Technol. 20 (2011) 108–115.
[3] N.P. Padture, M. Gell, E.H. Jordan, Thermal barrier coatings for gas
turbine engine applications, Science 296 (2002) 280–284.
[4] G. Di Girolamo, M. Alfano, L. Pagnotta, D. Sun, J. Zekonyte, R.J.
K. Wood, On the early stage isothermal oxidation of APS CoNiCrAlY
coatings, J. Mater. Eng. Perform. 21 (2012) 1989–1997.
[5] X.Q. Cao, R. Vassen, S. Stöver, Ceramic materials for thermal barrier
coatings, J. Eur. Ceram. Soc. 24 (2004) 1–10.
[6] G. Di Girolamo, L. Pagnotta, Thermally sprayed coatings for high
temperature applications, Rec. Patent on Mater. Sci. 4 (2011) 173–190.
[7] J.R. Davis, Handbook of Thermal Spray Technology, ASM International,
Materials Park, 2004.
[8] W.R. Chen, X. Wu, B.R. Marple, R.S. Lima, P.C. Patnaik, Pre-oxidation
and TGO growth behavior of an air-plasma-sprayed thermal barrier
coating, Surf. Coat. Technol. 202 (2008) 3787–3796.
[9] P. Puetz, X. Huang, R.S. Lima, Q. Yang, L. Zhao, Characterization of
transient oxide formation on CoNiCrAlY after heat treatment in vacuum
and air, Surf. Coat. Technol. 205 (2010) 647–657.
[10] M. Daroonparvar, M.A. Mat Yajid, N.M. Yusof, M.S. Hussain, Forma-
tion of a dense and continues Al2O3 layer in nano thermal barrier coating
systems for suppression of the spinels growth on the Al2O3 oxide scale
during oxidation, J. Alloy. Compd. 571 (2013) 205–220.
[11] M. Daroonparvar, M.A. Mat Yajid, N.M. Yusof, M.S. Hussain, Improved
thermally grown oxide scale in air plasma sprayed NiCrAlY/Nano-
YSZ coatings, J. Nanometer. 2013 (2013) 9 Article ID 520104
http://dx.doi.org/10.1155/2013/520104.
[12] R.S. Lima, B.R. Marple, Thermal spray coatings engineered from
nanostructured ceramic agglomerated powders for structural, thermal
barrier and biomedical applications: a review, J. Therm. Spray Technol.
16 (2007) 40–63.
[13] J. Wu, H.B. Guo, L. Zhou, L. Wang, S.K. Gong, Microstructure and
thermal properties of plasma sprayed thermal barrier coatings from
nanostructured YSZ, J. Therm. Spray Technol. 19 (2010) 1186–1194.
[14] R.S. Lima, A. Kucuk, C.C. Berndt, Bimodal distribution of mechanical
properties on plasma sprayed nanostructured partially stabilized zirconia,
Mater. Sci. Eng. A 327 (2002) 224–232.
[15] A.C. Fisher-Cripps, Nanoindentation, 3rd edition, Springer, New York,
2011.
[16] W.C. Oliver, G.M. Pharr, An improved technique for determining
hardness and elastic modulus using load and displacement sensing
indentation experiments, J. Mater. Res. 7 (1992) 1564–1583.
[17] W.C. Oliver, G.M. Pharr, Measurement of hardness and elastic modulus
by instrumented indentation: advances in understanding and refinements
to methodology, J. Mater. Res. 19 (2004) 3–20.
[18] W. Weibull, A statistical distribution function of wide applicability,
J. Appl. Mech. 18 (1951) 293–297.
[19] G. Di Girolamo, F. Marra, C. Blasi, E. Serra, T. Valente, Microstructure,
mechanical properties and thermal shock resistance of plasma sprayed
nanostructured zirconia coatings, Ceram. Int. 37 (2011) 2711–2717.
[20] L. Wang, Y. Wang, X.G. Sun, J.Q. He, Z.Y. Pan, C.H. Wang,
Microstructure and indentation mechanical properties of plasma sprayed
nano-bimodal and conventional ZrO2–8 wt%Y2O3 thermal barrier coat-
ings, Vacuum 86 (2012) 1174–1185.
[21] M. Saremi, A. Keyvanimahmoud, H. Sohi, Hot corrosion resistance and
mechanical behavior of atmospheric plasma sprayed conventional and
nanostructured zirconia coatings, Int. J. Mod. Phys. Conf. Ser. 5 (2012)
720–727.
[22] B. Liang, G. Zhang, H.L. Liao, C. Coddet, C.X. Ding, Structure and
tribological performance of nanostructured ZrO2–3 mol% Y2O3 coatings
deposited by air plasma spraying, J. Therm. Spray Technol. 19 (6) (2010)
1163–1170.